Detect positive and negative sources on an exposure and return a new table.SourceCatalog. More...

Inheritance diagram for lsst.meas.algorithms.detection.SourceDetectionTask:

Public Member Functions
def	__init__ (self, schema=None, kwds)
	Create the detection task. More...

def	run (self, table, exposure, doSmooth=True, sigma=None, clearMask=True, expId=None)

def	display (self, exposure, results, convolvedImage=None)

def	applyTempLocalBackground (self, exposure, middle, results)

def	clearMask (self, mask)

def	calculateKernelSize (self, sigma)

def	getPsf (self, exposure, sigma=None)

def	convolveImage (self, maskedImage, psf, doSmooth=True)

def	applyThreshold (self, middle, bbox, factor=1.0)

def	finalizeFootprints (self, mask, results, sigma, factor=1.0)

def	reEstimateBackground (self, maskedImage, backgrounds)

def	clearUnwantedResults (self, mask, results)

def	detectFootprints (self, exposure, doSmooth=True, sigma=None, clearMask=True, expId=None)

def	makeThreshold (self, image, thresholdParity, factor=1.0)

def	updatePeaks (self, fpSet, image, threshold)

def	tempWideBackgroundContext (self, exposure)

Static Public Member Functions
def	setEdgeBits (maskedImage, goodBBox, edgeBitmask)

Public Attributes
	negativeFlagKey

Static Public Attributes
	ConfigClass = SourceDetectionConfig

def	makeSourceCatalog = run
	An alias for run. More...

Detailed Description

Detect positive and negative sources on an exposure and return a new table.SourceCatalog.

Description

Detect positive and negative sources on an exposure and return a new table.SourceCatalog.

Task initialisation

Create the detection task. Most arguments are simply passed onto pipe.base.Task.

Parameters

schema	An lsst::afw::table::Schema used to create the output lsst.afw.table.SourceCatalog
**kwds	Keyword arguments passed to lsst.pipe.base.task.Task.__init__.

If schema is not None and configured for 'both' detections, a 'flags.negative' field will be added to label detections made with a negative threshold.

Note: This task can add fields to the schema, so any code calling this task must ensure that these columns are indeed present in the input match list; see Example

Invoking the Task

Run source detection and create a SourceCatalog of detections.

Parameters
----------
table : `lsst.afw.table.SourceTable`
    Table object that will be used to create the SourceCatalog.
exposure : `lsst.afw.image.Exposure`
    Exposure to process; DETECTED mask plane will be set in-place.
doSmooth : `bool`
    If True, smooth the image before detection using a Gaussian of width
    ``sigma``, or the measured PSF width. Set to False when running on
    e.g. a pre-convolved image, or a mask plane.
sigma : `float`
    Sigma of PSF (pixels); used for smoothing and to grow detections;
    if None then measure the sigma of the PSF of the exposure
clearMask : `bool`
    Clear DETECTED{,_NEGATIVE} planes before running detection.
expId : `int`
    Exposure identifier; unused by this implementation, but used for
    RNG seed by subclasses.

Returns
-------
result : `lsst.pipe.base.Struct`
  ``sources``
      The detected sources (`lsst.afw.table.SourceCatalog`)
  ``fpSets``
      The result resturned by `detectFootprints`
      (`lsst.pipe.base.Struct`).

Raises
------
ValueError
    If flags.negative is needed, but isn't in table's schema.
lsst.pipe.base.TaskError
    If sigma=None, doSmooth=True and the exposure has no PSF.

Notes
-----
If you want to avoid dealing with Sources and Tables, you can use
detectFootprints() to just get the `lsst.afw.detection.FootprintSet`s.

Configuration parameters

See SourceDetectionConfig

Debug variables

The command line task interface supports a flag -d to import debug.py from your PYTHONPATH; see Using lsstDebug to control debugging output for more about debug.py files.

The available variables in SourceDetectionTask are:

display

If True, display the exposure on afwDisplay.Display's frame 0. +ve detections in blue, -ve detections in cyan
If display > 1, display the convolved exposure on frame 1

A complete example of using SourceDetectionTask

This code is in measAlgTasks.py in the examples directory, and can be run as e.g.

examples/measAlgTasks.py --doDisplay

The example also runs the SingleFrameMeasurementTask; see meas_algorithms_measurement_Example for more explanation.

Import the task (there are some other standard imports; read the file if you're confused)

from lsst.meas.algorithms.detection import SourceDetectionTask

We need to create our task before processing any data as the task constructor can add an extra column to the schema, but first we need an almost-empty Schema

schema = afwTable.SourceTable.makeMinimalSchema()

after which we can call the constructor:

    config = SourceDetectionTask.ConfigClass()
    config.thresholdPolarity = "both"
    config.background.isNanSafe = True
    config.thresholdValue = 3
    detectionTask = SourceDetectionTask(config=config, schema=schema)

We're now ready to process the data (we could loop over multiple exposures/catalogues using the same task objects). First create the output table:

tab = afwTable.SourceTable.make(schema)

And process the image

result = detectionTask.run(tab, exposure)

(You may not be happy that the threshold was set in the config before creating the Task rather than being set separately for each exposure. You can reset it just before calling the run method if you must, but we should really implement a better solution).

We can then unpack and use the results:

    sources = result.sources
    print("Found %d sources (%d +ve, %d -ve)" % (len(sources), result.fpSets.numPos, result.fpSets.numNeg))

To investigate the Debug variables, put something like

import lsstDebug
def DebugInfo(name):
    di = lsstDebug.getInfo(name)        # N.b. lsstDebug.Info(name) would call us recursively
    if name == "lsst.meas.algorithms.detection":
        di.display = 1
    return di
lsstDebug.Info = DebugInfo

into your debug.py file and run measAlgTasks.py with the –debug flag.

Definition at line 168 of file detection.py.

Constructor & Destructor Documentation

◆ init()

def lsst.meas.algorithms.detection.SourceDetectionTask.__init__	(	self,
		schema = `None`,
		kwds
	)

Create the detection task.

Most arguments are simply passed onto pipe.base.Task.

Parameters

schema	An lsst::afw::table::Schema used to create the output lsst.afw.table.SourceCatalog
**kwds	Keyword arguments passed to lsst.pipe.base.task.Task.__init__.

If schema is not None and configured for 'both' detections, a 'flags.negative' field will be added to label detections made with a negative threshold.

Note: This task can add fields to the schema, so any code calling this task must ensure that these columns are indeed present in the input match list; see Example

Definition at line 265 of file detection.py.

     def __init__(self, schema=None, **kwds):
         """!Create the detection task.  Most arguments are simply passed onto pipe.base.Task.
 
         @param schema An lsst::afw::table::Schema used to create the output lsst.afw.table.SourceCatalog
         @param **kwds Keyword arguments passed to lsst.pipe.base.task.Task.__init__.
 
         If schema is not None and configured for 'both' detections,
         a 'flags.negative' field will be added to label detections made with a
         negative threshold.
 
         @note This task can add fields to the schema, so any code calling this task must ensure that
         these columns are indeed present in the input match list; see @ref Example
         """
         pipeBase.Task.__init__(self, **kwds)
         if schema is not None and self.config.thresholdPolarity == "both":
             self.negativeFlagKey = schema.addField(
                 "flags_negative", type="Flag",
                 doc="set if source was detected as significantly negative"
             )
         else:
             if self.config.thresholdPolarity == "both":
                 self.log.warn("Detection polarity set to 'both', but no flag will be "
                               "set to distinguish between positive and negative detections")
             self.negativeFlagKey = None
         if self.config.reEstimateBackground:
             self.makeSubtask("background")
         if self.config.doTempLocalBackground:
             self.makeSubtask("tempLocalBackground")
         if self.config.doTempWideBackground:
             self.makeSubtask("tempWideBackground")
 

Member Function Documentation

◆ applyTempLocalBackground()

def lsst.meas.algorithms.detection.SourceDetectionTask.applyTempLocalBackground	(	self,
		exposure,
		middle,
		results
	)

Apply a temporary local background subtraction

This temporary local background serves to suppress noise fluctuations
in the wings of bright objects.

Peaks in the footprints will be updated.

Parameters
----------
exposure : `lsst.afw.image.Exposure`
    Exposure for which to fit local background.
middle : `lsst.afw.image.MaskedImage`
    Convolved image on which detection will be performed
    (typically smaller than ``exposure`` because the
    half-kernel has been removed around the edges).
results : `lsst.pipe.base.Struct`
    Results of the 'detectFootprints' method, containing positive and
    negative footprints (which contain the peak positions that we will
    plot). This is a `Struct` with ``positive`` and ``negative``
    elements that are of type `lsst.afw.detection.FootprintSet`.

Definition at line 414 of file detection.py.

     def applyTempLocalBackground(self, exposure, middle, results):
         """Apply a temporary local background subtraction
 
         This temporary local background serves to suppress noise fluctuations
         in the wings of bright objects.
 
         Peaks in the footprints will be updated.
 
         Parameters
         ----------
         exposure : `lsst.afw.image.Exposure`
             Exposure for which to fit local background.
         middle : `lsst.afw.image.MaskedImage`
             Convolved image on which detection will be performed
             (typically smaller than ``exposure`` because the
             half-kernel has been removed around the edges).
         results : `lsst.pipe.base.Struct`
             Results of the 'detectFootprints' method, containing positive and
             negative footprints (which contain the peak positions that we will
             plot). This is a `Struct` with ``positive`` and ``negative``
             elements that are of type `lsst.afw.detection.FootprintSet`.
         """
         # Subtract the local background from the smoothed image. Since we
         # never use the smoothed again we don't need to worry about adding
         # it back in.
         bg = self.tempLocalBackground.fitBackground(exposure.getMaskedImage())
         bgImage = bg.getImageF()
         middle -= bgImage.Factory(bgImage, middle.getBBox())
         thresholdPos = self.makeThreshold(middle, "positive")
         thresholdNeg = self.makeThreshold(middle, "negative")
         if self.config.thresholdPolarity != "negative":
             self.updatePeaks(results.positive, middle, thresholdPos)
         if self.config.thresholdPolarity != "positive":
             self.updatePeaks(results.negative, middle, thresholdNeg)
 

◆ applyThreshold()

def lsst.meas.algorithms.detection.SourceDetectionTask.applyThreshold	(	self,
		middle,
		bbox,
		factor = `1.0`
	)

Apply thresholds to the convolved image

Identifies ``Footprint``s, both positive and negative.

The threshold can be modified by the provided multiplication
``factor``.

Parameters
----------
middle : `lsst.afw.image.MaskedImage`
    Convolved image to threshold.
bbox : `lsst.geom.Box2I`
    Bounding box of unconvolved image.
factor : `float`
    Multiplier for the configured threshold.

Return Struct contents
----------------------
positive : `lsst.afw.detection.FootprintSet` or `None`
    Positive detection footprints, if configured.
negative : `lsst.afw.detection.FootprintSet` or `None`
    Negative detection footprints, if configured.
factor : `float`
    Multiplier for the configured threshold.

Definition at line 570 of file detection.py.

     def applyThreshold(self, middle, bbox, factor=1.0):
         """Apply thresholds to the convolved image
 
         Identifies ``Footprint``s, both positive and negative.
 
         The threshold can be modified by the provided multiplication
         ``factor``.
 
         Parameters
         ----------
         middle : `lsst.afw.image.MaskedImage`
             Convolved image to threshold.
         bbox : `lsst.geom.Box2I`
             Bounding box of unconvolved image.
         factor : `float`
             Multiplier for the configured threshold.
 
         Return Struct contents
         ----------------------
         positive : `lsst.afw.detection.FootprintSet` or `None`
             Positive detection footprints, if configured.
         negative : `lsst.afw.detection.FootprintSet` or `None`
             Negative detection footprints, if configured.
         factor : `float`
             Multiplier for the configured threshold.
         """
         results = pipeBase.Struct(positive=None, negative=None, factor=factor)
         # Detect the Footprints (peaks may be replaced if doTempLocalBackground)
         if self.config.reEstimateBackground or self.config.thresholdPolarity != "negative":
             threshold = self.makeThreshold(middle, "positive", factor=factor)
             results.positive = afwDet.FootprintSet(
                 middle,
                 threshold,
                 "DETECTED",
                 self.config.minPixels
             )
             results.positive.setRegion(bbox)
         if self.config.reEstimateBackground or self.config.thresholdPolarity != "positive":
             threshold = self.makeThreshold(middle, "negative", factor=factor)
             results.negative = afwDet.FootprintSet(
                 middle,
                 threshold,
                 "DETECTED_NEGATIVE",
                 self.config.minPixels
             )
             results.negative.setRegion(bbox)
 
         return results
 

◆ calculateKernelSize()

def lsst.meas.algorithms.detection.SourceDetectionTask.calculateKernelSize	(	self,
		sigma
	)

Calculate size of smoothing kernel

Uses the ``nSigmaForKernel`` configuration parameter. Note
that that is the full width of the kernel bounding box
(so a value of 7 means 3.5 sigma on either side of center).
The value will be rounded up to the nearest odd integer.

Parameters
----------
sigma : `float`
    Gaussian sigma of smoothing kernel.

Returns
-------
size : `int`
    Size of the smoothing kernel.

Definition at line 462 of file detection.py.

     def calculateKernelSize(self, sigma):
         """Calculate size of smoothing kernel
 
         Uses the ``nSigmaForKernel`` configuration parameter. Note
         that that is the full width of the kernel bounding box
         (so a value of 7 means 3.5 sigma on either side of center).
         The value will be rounded up to the nearest odd integer.
 
         Parameters
         ----------
         sigma : `float`
             Gaussian sigma of smoothing kernel.
 
         Returns
         -------
         size : `int`
             Size of the smoothing kernel.
         """
         return (int(sigma * self.config.nSigmaForKernel + 0.5)//2)*2 + 1  # make sure it is odd
 

◆ clearMask()

def lsst.meas.algorithms.detection.SourceDetectionTask.clearMask	(	self,
		mask
	)

Clear the DETECTED and DETECTED_NEGATIVE mask planes

Removes any previous detection mask in preparation for a new
detection pass.

Parameters
----------
mask : `lsst.afw.image.Mask`
    Mask to be cleared.

Definition at line 449 of file detection.py.

     def clearMask(self, mask):
         """Clear the DETECTED and DETECTED_NEGATIVE mask planes
 
         Removes any previous detection mask in preparation for a new
         detection pass.
 
         Parameters
         ----------
         mask : `lsst.afw.image.Mask`
             Mask to be cleared.
         """
         mask &= ~(mask.getPlaneBitMask("DETECTED") | mask.getPlaneBitMask("DETECTED_NEGATIVE"))
 

◆ clearUnwantedResults()

def lsst.meas.algorithms.detection.SourceDetectionTask.clearUnwantedResults	(	self,
		mask,
		results
	)

Clear unwanted results from the Struct of results

If we specifically want only positive or only negative detections,
drop the ones we don't want, and its associated mask plane.

Parameters
----------
mask : `lsst.afw.image.Mask`
    Mask image.
results : `lsst.pipe.base.Struct`
    Detection results, with ``positive`` and ``negative`` elements;
    modified.

Definition at line 705 of file detection.py.

     def clearUnwantedResults(self, mask, results):
         """Clear unwanted results from the Struct of results
 
         If we specifically want only positive or only negative detections,
         drop the ones we don't want, and its associated mask plane.
 
         Parameters
         ----------
         mask : `lsst.afw.image.Mask`
             Mask image.
         results : `lsst.pipe.base.Struct`
             Detection results, with ``positive`` and ``negative`` elements;
             modified.
         """
         if self.config.thresholdPolarity == "positive":
             if self.config.reEstimateBackground:
                 mask &= ~mask.getPlaneBitMask("DETECTED_NEGATIVE")
             results.negative = None
         elif self.config.thresholdPolarity == "negative":
             if self.config.reEstimateBackground:
                 mask &= ~mask.getPlaneBitMask("DETECTED")
             results.positive = None
 

◆ convolveImage()

def lsst.meas.algorithms.detection.SourceDetectionTask.convolveImage	(	self,
		maskedImage,
		psf,
		doSmooth = `True`
	)

Convolve the image with the PSF

We convolve the image with a Gaussian approximation to the PSF,
because this is separable and therefore fast. It's technically a
correlation rather than a convolution, but since we use a symmetric
Gaussian there's no difference.

The convolution can be disabled with ``doSmooth=False``. If we do
convolve, we mask the edges as ``EDGE`` and return the convolved image
with the edges removed. This is because we can't convolve the edges
because the kernel would extend off the image.

Parameters
----------
maskedImage : `lsst.afw.image.MaskedImage`
    Image to convolve.
psf : `lsst.afw.detection.Psf`
    PSF to convolve with (actually with a Gaussian approximation
    to it).
doSmooth : `bool`
    Actually do the convolution? Set to False when running on
    e.g. a pre-convolved image, or a mask plane.

Return Struct contents
----------------------
middle : `lsst.afw.image.MaskedImage`
    Convolved image, without the edges.
sigma : `float`
    Gaussian sigma used for the convolution.

Definition at line 509 of file detection.py.

     def convolveImage(self, maskedImage, psf, doSmooth=True):
         """Convolve the image with the PSF
 
         We convolve the image with a Gaussian approximation to the PSF,
         because this is separable and therefore fast. It's technically a
         correlation rather than a convolution, but since we use a symmetric
         Gaussian there's no difference.
 
         The convolution can be disabled with ``doSmooth=False``. If we do
         convolve, we mask the edges as ``EDGE`` and return the convolved image
         with the edges removed. This is because we can't convolve the edges
         because the kernel would extend off the image.
 
         Parameters
         ----------
         maskedImage : `lsst.afw.image.MaskedImage`
             Image to convolve.
         psf : `lsst.afw.detection.Psf`
             PSF to convolve with (actually with a Gaussian approximation
             to it).
         doSmooth : `bool`
             Actually do the convolution? Set to False when running on
             e.g. a pre-convolved image, or a mask plane.
 
         Return Struct contents
         ----------------------
         middle : `lsst.afw.image.MaskedImage`
             Convolved image, without the edges.
         sigma : `float`
             Gaussian sigma used for the convolution.
         """
         self.metadata.set("doSmooth", doSmooth)
         sigma = psf.computeShape().getDeterminantRadius()
         self.metadata.set("sigma", sigma)
 
         if not doSmooth:
             middle = maskedImage.Factory(maskedImage)
             return pipeBase.Struct(middle=middle, sigma=sigma)
 
         # Smooth using a Gaussian (which is separable, hence fast) of width sigma
         # Make a SingleGaussian (separable) kernel with the 'sigma'
         kWidth = self.calculateKernelSize(sigma)
         self.metadata.set("smoothingKernelWidth", kWidth)
         gaussFunc = afwMath.GaussianFunction1D(sigma)
         gaussKernel = afwMath.SeparableKernel(kWidth, kWidth, gaussFunc, gaussFunc)
 
         convolvedImage = maskedImage.Factory(maskedImage.getBBox())
 
         afwMath.convolve(convolvedImage, maskedImage, gaussKernel, afwMath.ConvolutionControl())
         #
         # Only search psf-smoothed part of frame
         #
         goodBBox = gaussKernel.shrinkBBox(convolvedImage.getBBox())
         middle = convolvedImage.Factory(convolvedImage, goodBBox, afwImage.PARENT, False)
         #
         # Mark the parts of the image outside goodBBox as EDGE
         #
         self.setEdgeBits(maskedImage, goodBBox, maskedImage.getMask().getPlaneBitMask("EDGE"))
 
         return pipeBase.Struct(middle=middle, sigma=sigma)
 

◆ detectFootprints()

def lsst.meas.algorithms.detection.SourceDetectionTask.detectFootprints	(	self,
		exposure,
		doSmooth = `True`,
		sigma = `None`,
		clearMask = `True`,
		expId = `None`
	)

Detect footprints on an exposure.

Parameters
----------
exposure : `lsst.afw.image.Exposure`
    Exposure to process; DETECTED{,_NEGATIVE} mask plane will be
    set in-place.
doSmooth : `bool`, optional
    If True, smooth the image before detection using a Gaussian
    of width ``sigma``, or the measured PSF width of ``exposure``.
    Set to False when running on e.g. a pre-convolved image, or a mask
    plane.
sigma : `float`, optional
    Gaussian Sigma of PSF (pixels); used for smoothing and to grow
    detections; if `None` then measure the sigma of the PSF of the
    ``exposure``.
clearMask : `bool`, optional
    Clear both DETECTED and DETECTED_NEGATIVE planes before running
    detection.
expId : `dict`, optional
    Exposure identifier; unused by this implementation, but used for
    RNG seed by subclasses.

Return Struct contents
----------------------
positive : `lsst.afw.detection.FootprintSet`
    Positive polarity footprints (may be `None`)
negative : `lsst.afw.detection.FootprintSet`
    Negative polarity footprints (may be `None`)
numPos : `int`
    Number of footprints in positive or 0 if detection polarity was
    negative.
numNeg : `int`
    Number of footprints in negative or 0 if detection polarity was
    positive.
background : `lsst.afw.math.BackgroundList`
    Re-estimated background.  `None` if
    ``reEstimateBackground==False``.
factor : `float`
    Multiplication factor applied to the configured detection
    threshold.

Definition at line 729 of file detection.py.

     def detectFootprints(self, exposure, doSmooth=True, sigma=None, clearMask=True, expId=None):
         """Detect footprints on an exposure.
 
         Parameters
         ----------
         exposure : `lsst.afw.image.Exposure`
             Exposure to process; DETECTED{,_NEGATIVE} mask plane will be
             set in-place.
         doSmooth : `bool`, optional
             If True, smooth the image before detection using a Gaussian
             of width ``sigma``, or the measured PSF width of ``exposure``.
             Set to False when running on e.g. a pre-convolved image, or a mask
             plane.
         sigma : `float`, optional
             Gaussian Sigma of PSF (pixels); used for smoothing and to grow
             detections; if `None` then measure the sigma of the PSF of the
             ``exposure``.
         clearMask : `bool`, optional
             Clear both DETECTED and DETECTED_NEGATIVE planes before running
             detection.
         expId : `dict`, optional
             Exposure identifier; unused by this implementation, but used for
             RNG seed by subclasses.
 
         Return Struct contents
         ----------------------
         positive : `lsst.afw.detection.FootprintSet`
             Positive polarity footprints (may be `None`)
         negative : `lsst.afw.detection.FootprintSet`
             Negative polarity footprints (may be `None`)
         numPos : `int`
             Number of footprints in positive or 0 if detection polarity was
             negative.
         numNeg : `int`
             Number of footprints in negative or 0 if detection polarity was
             positive.
         background : `lsst.afw.math.BackgroundList`
             Re-estimated background.  `None` if
             ``reEstimateBackground==False``.
         factor : `float`
             Multiplication factor applied to the configured detection
             threshold.
         """
         maskedImage = exposure.maskedImage
 
         if clearMask:
             self.clearMask(maskedImage.getMask())
 
         psf = self.getPsf(exposure, sigma=sigma)
         with self.tempWideBackgroundContext(exposure):
             convolveResults = self.convolveImage(maskedImage, psf, doSmooth=doSmooth)
             middle = convolveResults.middle
             sigma = convolveResults.sigma
 
             results = self.applyThreshold(middle, maskedImage.getBBox())
             results.background = afwMath.BackgroundList()
             if self.config.doTempLocalBackground:
                 self.applyTempLocalBackground(exposure, middle, results)
             self.finalizeFootprints(maskedImage.mask, results, sigma)
 
             if self.config.reEstimateBackground:
                 self.reEstimateBackground(maskedImage, results.background)
 
             self.clearUnwantedResults(maskedImage.getMask(), results)
             self.display(exposure, results, middle)
 
         return results
 

◆ display()

def lsst.meas.algorithms.detection.SourceDetectionTask.display	(	self,
		exposure,
		results,
		convolvedImage = `None`
	)

Display detections if so configured

Displays the ``exposure`` in frame 0, overlays the detection peaks.

Requires that ``lsstDebug`` has been set up correctly, so that
``lsstDebug.Info("lsst.meas.algorithms.detection")`` evaluates `True`.

If the ``convolvedImage`` is non-`None` and
``lsstDebug.Info("lsst.meas.algorithms.detection") > 1``, the
``convolvedImage`` will be displayed in frame 1.

Parameters
----------
exposure : `lsst.afw.image.Exposure`
    Exposure to display, on which will be plotted the detections.
results : `lsst.pipe.base.Struct`
    Results of the 'detectFootprints' method, containing positive and
    negative footprints (which contain the peak positions that we will
    plot). This is a `Struct` with ``positive`` and ``negative``
    elements that are of type `lsst.afw.detection.FootprintSet`.
convolvedImage : `lsst.afw.image.Image`, optional
    Convolved image used for thresholding.

Definition at line 360 of file detection.py.

     def display(self, exposure, results, convolvedImage=None):
         """Display detections if so configured
 
         Displays the ``exposure`` in frame 0, overlays the detection peaks.
 
         Requires that ``lsstDebug`` has been set up correctly, so that
         ``lsstDebug.Info("lsst.meas.algorithms.detection")`` evaluates `True`.
 
         If the ``convolvedImage`` is non-`None` and
         ``lsstDebug.Info("lsst.meas.algorithms.detection") > 1``, the
         ``convolvedImage`` will be displayed in frame 1.
 
         Parameters
         ----------
         exposure : `lsst.afw.image.Exposure`
             Exposure to display, on which will be plotted the detections.
         results : `lsst.pipe.base.Struct`
             Results of the 'detectFootprints' method, containing positive and
             negative footprints (which contain the peak positions that we will
             plot). This is a `Struct` with ``positive`` and ``negative``
             elements that are of type `lsst.afw.detection.FootprintSet`.
         convolvedImage : `lsst.afw.image.Image`, optional
             Convolved image used for thresholding.
         """
         try:
             import lsstDebug
             display = lsstDebug.Info(__name__).display
         except ImportError:
             try:
                 display
             except NameError:
                 display = False
         if not display:
             return
 
         afwDisplay.setDefaultMaskTransparency(75)
 
         disp0 = afwDisplay.Display(frame=0)
         disp0.mtv(exposure, title="detection")
 
         def plotPeaks(fps, ctype):
             if fps is None:
                 return
             with disp0.Buffering():
                 for fp in fps.getFootprints():
                     for pp in fp.getPeaks():
                         disp0.dot("+", pp.getFx(), pp.getFy(), ctype=ctype)
         plotPeaks(results.positive, "yellow")
         plotPeaks(results.negative, "red")
 
         if convolvedImage and display > 1:
             disp1 = afwDisplay.Display(frame=1)
             disp1.mtv(convolvedImage, title="PSF smoothed")
 

◆ finalizeFootprints()

def lsst.meas.algorithms.detection.SourceDetectionTask.finalizeFootprints	(	self,
		mask,
		results,
		sigma,
		factor = `1.0`
	)

Finalize the detected footprints

Grows the footprints, sets the ``DETECTED`` and ``DETECTED_NEGATIVE``
mask planes, and logs the results.

``numPos`` (number of positive footprints), ``numPosPeaks`` (number
of positive peaks), ``numNeg`` (number of negative footprints),
``numNegPeaks`` (number of negative peaks) entries are added to the
detection results.

Parameters
----------
mask : `lsst.afw.image.Mask`
    Mask image on which to flag detected pixels.
results : `lsst.pipe.base.Struct`
    Struct of detection results, including ``positive`` and
    ``negative`` entries; modified.
sigma : `float`
    Gaussian sigma of PSF.
factor : `float`
    Multiplier for the configured threshold.

Definition at line 619 of file detection.py.

     def finalizeFootprints(self, mask, results, sigma, factor=1.0):
         """Finalize the detected footprints
 
         Grows the footprints, sets the ``DETECTED`` and ``DETECTED_NEGATIVE``
         mask planes, and logs the results.
 
         ``numPos`` (number of positive footprints), ``numPosPeaks`` (number
         of positive peaks), ``numNeg`` (number of negative footprints),
         ``numNegPeaks`` (number of negative peaks) entries are added to the
         detection results.
 
         Parameters
         ----------
         mask : `lsst.afw.image.Mask`
             Mask image on which to flag detected pixels.
         results : `lsst.pipe.base.Struct`
             Struct of detection results, including ``positive`` and
             ``negative`` entries; modified.
         sigma : `float`
             Gaussian sigma of PSF.
         factor : `float`
             Multiplier for the configured threshold.
         """
         for polarity, maskName in (("positive", "DETECTED"), ("negative", "DETECTED_NEGATIVE")):
             fpSet = getattr(results, polarity)
             if fpSet is None:
                 continue
             if self.config.nSigmaToGrow > 0:
                 nGrow = int((self.config.nSigmaToGrow * sigma) + 0.5)
                 self.metadata.set("nGrow", nGrow)
                 if self.config.combinedGrow:
                     fpSet = afwDet.FootprintSet(fpSet, nGrow, self.config.isotropicGrow)
                 else:
                     stencil = (afwGeom.Stencil.CIRCLE if self.config.isotropicGrow else
                                afwGeom.Stencil.MANHATTAN)
                     for fp in fpSet:
                         fp.dilate(nGrow, stencil)
             fpSet.setMask(mask, maskName)
             if not self.config.returnOriginalFootprints:
                 setattr(results, polarity, fpSet)
 
         results.numPos = 0
         results.numPosPeaks = 0
         results.numNeg = 0
         results.numNegPeaks = 0
         positive = ""
         negative = ""
 
         if results.positive is not None:
             results.numPos = len(results.positive.getFootprints())
             results.numPosPeaks = sum(len(fp.getPeaks()) for fp in results.positive.getFootprints())
             positive = " %d positive peaks in %d footprints" % (results.numPosPeaks, results.numPos)
         if results.negative is not None:
             results.numNeg = len(results.negative.getFootprints())
             results.numNegPeaks = sum(len(fp.getPeaks()) for fp in results.negative.getFootprints())
             negative = " %d negative peaks in %d footprints" % (results.numNegPeaks, results.numNeg)
 
         self.log.info("Detected%s%s%s to %g %s" %
                       (positive, " and" if positive and negative else "", negative,
                        self.config.thresholdValue*self.config.includeThresholdMultiplier*factor,
                        "DN" if self.config.thresholdType == "value" else "sigma"))
 

◆ getPsf()

def lsst.meas.algorithms.detection.SourceDetectionTask.getPsf	(	self,
		exposure,
		sigma = `None`
	)

Retrieve the PSF for an exposure

If ``sigma`` is provided, we make a ``GaussianPsf`` with that,
otherwise use the one from the ``exposure``.

Parameters
----------
exposure : `lsst.afw.image.Exposure`
    Exposure from which to retrieve the PSF.
sigma : `float`, optional
    Gaussian sigma to use if provided.

Returns
-------
psf : `lsst.afw.detection.Psf`
    PSF to use for detection.

Definition at line 482 of file detection.py.

     def getPsf(self, exposure, sigma=None):
         """Retrieve the PSF for an exposure
 
         If ``sigma`` is provided, we make a ``GaussianPsf`` with that,
         otherwise use the one from the ``exposure``.
 
         Parameters
         ----------
         exposure : `lsst.afw.image.Exposure`
             Exposure from which to retrieve the PSF.
         sigma : `float`, optional
             Gaussian sigma to use if provided.
 
         Returns
         -------
         psf : `lsst.afw.detection.Psf`
             PSF to use for detection.
         """
         if sigma is None:
             psf = exposure.getPsf()
             if psf is None:
                 raise RuntimeError("Unable to determine PSF to use for detection: no sigma provided")
             sigma = psf.computeShape().getDeterminantRadius()
         size = self.calculateKernelSize(sigma)
         psf = afwDet.GaussianPsf(size, size, sigma)
         return psf
 

◆ makeThreshold()

def lsst.meas.algorithms.detection.SourceDetectionTask.makeThreshold	(	self,
		image,
		thresholdParity,
		factor = `1.0`
	)

Make an afw.detection.Threshold object corresponding to the task's
configuration and the statistics of the given image.

Parameters
----------
image : `afw.image.MaskedImage`
    Image to measure noise statistics from if needed.
thresholdParity: `str`
    One of "positive" or "negative", to set the kind of fluctuations
    the Threshold will detect.
factor : `float`
    Factor by which to multiply the configured detection threshold.
    This is useful for tweaking the detection threshold slightly.

Returns
-------
threshold : `lsst.afw.detection.Threshold`
    Detection threshold.

Definition at line 797 of file detection.py.

     def makeThreshold(self, image, thresholdParity, factor=1.0):
         """Make an afw.detection.Threshold object corresponding to the task's
         configuration and the statistics of the given image.
 
         Parameters
         ----------
         image : `afw.image.MaskedImage`
             Image to measure noise statistics from if needed.
         thresholdParity: `str`
             One of "positive" or "negative", to set the kind of fluctuations
             the Threshold will detect.
         factor : `float`
             Factor by which to multiply the configured detection threshold.
             This is useful for tweaking the detection threshold slightly.
 
         Returns
         -------
         threshold : `lsst.afw.detection.Threshold`
             Detection threshold.
         """
         parity = False if thresholdParity == "negative" else True
         thresholdValue = self.config.thresholdValue
         thresholdType = self.config.thresholdType
         if self.config.thresholdType == 'stdev':
             bad = image.getMask().getPlaneBitMask(self.config.statsMask)
             sctrl = afwMath.StatisticsControl()
             sctrl.setAndMask(bad)
             stats = afwMath.makeStatistics(image, afwMath.STDEVCLIP, sctrl)
             thresholdValue *= stats.getValue(afwMath.STDEVCLIP)
             thresholdType = 'value'
 
         threshold = afwDet.createThreshold(thresholdValue*factor, thresholdType, parity)
         threshold.setIncludeMultiplier(self.config.includeThresholdMultiplier)
         return threshold
 

◆ reEstimateBackground()

def lsst.meas.algorithms.detection.SourceDetectionTask.reEstimateBackground	(	self,
		maskedImage,
		backgrounds
	)

Estimate the background after detection

Parameters
----------
maskedImage : `lsst.afw.image.MaskedImage`
    Image on which to estimate the background.
backgrounds : `lsst.afw.math.BackgroundList`
    List of backgrounds; modified.

Returns
-------
bg : `lsst.afw.math.backgroundMI`
    Empirical background model.

Definition at line 681 of file detection.py.

     def reEstimateBackground(self, maskedImage, backgrounds):
         """Estimate the background after detection
 
         Parameters
         ----------
         maskedImage : `lsst.afw.image.MaskedImage`
             Image on which to estimate the background.
         backgrounds : `lsst.afw.math.BackgroundList`
             List of backgrounds; modified.
 
         Returns
         -------
         bg : `lsst.afw.math.backgroundMI`
             Empirical background model.
         """
         bg = self.background.fitBackground(maskedImage)
         if self.config.adjustBackground:
             self.log.warn("Fiddling the background by %g", self.config.adjustBackground)
             bg += self.config.adjustBackground
         self.log.info("Resubtracting the background after object detection")
         maskedImage -= bg.getImageF()
         backgrounds.append(bg)
         return bg
 

◆ run()

def lsst.meas.algorithms.detection.SourceDetectionTask.run	(	self,
		table,
		exposure,
		doSmooth = `True`,
		sigma = `None`,
		clearMask = `True`,
		expId = `None`
	)

Run source detection and create a SourceCatalog of detections.

Parameters
----------
table : `lsst.afw.table.SourceTable`
    Table object that will be used to create the SourceCatalog.
exposure : `lsst.afw.image.Exposure`
    Exposure to process; DETECTED mask plane will be set in-place.
doSmooth : `bool`
    If True, smooth the image before detection using a Gaussian of width
    ``sigma``, or the measured PSF width. Set to False when running on
    e.g. a pre-convolved image, or a mask plane.
sigma : `float`
    Sigma of PSF (pixels); used for smoothing and to grow detections;
    if None then measure the sigma of the PSF of the exposure
clearMask : `bool`
    Clear DETECTED{,_NEGATIVE} planes before running detection.
expId : `int`
    Exposure identifier; unused by this implementation, but used for
    RNG seed by subclasses.

Returns
-------
result : `lsst.pipe.base.Struct`
  ``sources``
      The detected sources (`lsst.afw.table.SourceCatalog`)
  ``fpSets``
      The result resturned by `detectFootprints`
      (`lsst.pipe.base.Struct`).

Raises
------
ValueError
    If flags.negative is needed, but isn't in table's schema.
lsst.pipe.base.TaskError
    If sigma=None, doSmooth=True and the exposure has no PSF.

Notes
-----
If you want to avoid dealing with Sources and Tables, you can use
detectFootprints() to just get the `lsst.afw.detection.FootprintSet`s.

Definition at line 297 of file detection.py.

     def run(self, table, exposure, doSmooth=True, sigma=None, clearMask=True, expId=None):
         """Run source detection and create a SourceCatalog of detections.
 
         Parameters
         ----------
         table : `lsst.afw.table.SourceTable`
             Table object that will be used to create the SourceCatalog.
         exposure : `lsst.afw.image.Exposure`
             Exposure to process; DETECTED mask plane will be set in-place.
         doSmooth : `bool`
             If True, smooth the image before detection using a Gaussian of width
             ``sigma``, or the measured PSF width. Set to False when running on
             e.g. a pre-convolved image, or a mask plane.
         sigma : `float`
             Sigma of PSF (pixels); used for smoothing and to grow detections;
             if None then measure the sigma of the PSF of the exposure
         clearMask : `bool`
             Clear DETECTED{,_NEGATIVE} planes before running detection.
         expId : `int`
             Exposure identifier; unused by this implementation, but used for
             RNG seed by subclasses.
 
         Returns
         -------
         result : `lsst.pipe.base.Struct`
           ``sources``
               The detected sources (`lsst.afw.table.SourceCatalog`)
           ``fpSets``
               The result resturned by `detectFootprints`
               (`lsst.pipe.base.Struct`).
 
         Raises
         ------
         ValueError
             If flags.negative is needed, but isn't in table's schema.
         lsst.pipe.base.TaskError
             If sigma=None, doSmooth=True and the exposure has no PSF.
 
         Notes
         -----
         If you want to avoid dealing with Sources and Tables, you can use
         detectFootprints() to just get the `lsst.afw.detection.FootprintSet`s.
         """
         if self.negativeFlagKey is not None and self.negativeFlagKey not in table.getSchema():
             raise ValueError("Table has incorrect Schema")
         results = self.detectFootprints(exposure=exposure, doSmooth=doSmooth, sigma=sigma,
                                         clearMask=clearMask, expId=expId)
         sources = afwTable.SourceCatalog(table)
         sources.reserve(results.numPos + results.numNeg)
         if results.negative:
             results.negative.makeSources(sources)
             if self.negativeFlagKey:
                 for record in sources:
                     record.set(self.negativeFlagKey, True)
         if results.positive:
             results.positive.makeSources(sources)
         results.fpSets = results.copy()  # Backward compatibility
         results.sources = sources
         return results
 

◆ setEdgeBits()

def lsst.meas.algorithms.detection.SourceDetectionTask.setEdgeBits	(	maskedImage,
		goodBBox,
		edgeBitmask
	)

static

Set the edgeBitmask bits for all of maskedImage outside goodBBox

Parameters
----------
maskedImage : `lsst.afw.image.MaskedImage`
    Image on which to set edge bits in the mask.
goodBBox : `lsst.geom.Box2I`
    Bounding box of good pixels, in ``LOCAL`` coordinates.
edgeBitmask : `lsst.afw.image.MaskPixel`
    Bit mask to OR with the existing mask bits in the region
    outside ``goodBBox``.

Definition at line 875 of file detection.py.

     def setEdgeBits(maskedImage, goodBBox, edgeBitmask):
         """Set the edgeBitmask bits for all of maskedImage outside goodBBox
 
         Parameters
         ----------
         maskedImage : `lsst.afw.image.MaskedImage`
             Image on which to set edge bits in the mask.
         goodBBox : `lsst.geom.Box2I`
             Bounding box of good pixels, in ``LOCAL`` coordinates.
         edgeBitmask : `lsst.afw.image.MaskPixel`
             Bit mask to OR with the existing mask bits in the region
             outside ``goodBBox``.
         """
         msk = maskedImage.getMask()
 
         mx0, my0 = maskedImage.getXY0()
         for x0, y0, w, h in ([0, 0,
                               msk.getWidth(), goodBBox.getBeginY() - my0],
                              [0, goodBBox.getEndY() - my0, msk.getWidth(),
                               maskedImage.getHeight() - (goodBBox.getEndY() - my0)],
                              [0, 0,
                               goodBBox.getBeginX() - mx0, msk.getHeight()],
                              [goodBBox.getEndX() - mx0, 0,
                               maskedImage.getWidth() - (goodBBox.getEndX() - mx0), msk.getHeight()],
                              ):
             edgeMask = msk.Factory(msk, lsst.geom.BoxI(lsst.geom.PointI(x0, y0),
                                                        lsst.geom.ExtentI(w, h)), afwImage.LOCAL)
             edgeMask |= edgeBitmask
 

◆ tempWideBackgroundContext()

def lsst.meas.algorithms.detection.SourceDetectionTask.tempWideBackgroundContext	(	self,
		exposure
	)

Context manager for removing wide (large-scale) background

Removing a wide (large-scale) background helps to suppress the
detection of large footprints that may overwhelm the deblender.
It does, however, set a limit on the maximum scale of objects.

The background that we remove will be restored upon exit from
the context manager.

Parameters
----------
exposure : `lsst.afw.image.Exposure`
    Exposure on which to remove large-scale background.

Returns
-------
context : context manager
    Context manager that will ensure the temporary wide background
    is restored.

Definition at line 905 of file detection.py.

     def tempWideBackgroundContext(self, exposure):
         """Context manager for removing wide (large-scale) background
 
         Removing a wide (large-scale) background helps to suppress the
         detection of large footprints that may overwhelm the deblender.
         It does, however, set a limit on the maximum scale of objects.
 
         The background that we remove will be restored upon exit from
         the context manager.
 
         Parameters
         ----------
         exposure : `lsst.afw.image.Exposure`
             Exposure on which to remove large-scale background.
 
         Returns
         -------
         context : context manager
             Context manager that will ensure the temporary wide background
             is restored.
         """
         doTempWideBackground = self.config.doTempWideBackground
         if doTempWideBackground:
             self.log.info("Applying temporary wide background subtraction")
             original = exposure.maskedImage.image.array[:].copy()
             self.tempWideBackground.run(exposure).background
             # Remove NO_DATA regions (e.g., edge of the field-of-view); these can cause detections after
             # subtraction because of extrapolation of the background model into areas with no constraints.
             image = exposure.maskedImage.image
             mask = exposure.maskedImage.mask
             noData = mask.array & mask.getPlaneBitMask("NO_DATA") > 0
             isGood = mask.array & mask.getPlaneBitMask(self.config.statsMask) == 0
             image.array[noData] = np.median(image.array[~noData & isGood])
         try:
             yield
         finally:
             if doTempWideBackground:
                 exposure.maskedImage.image.array[:] = original
 
 

◆ updatePeaks()

def lsst.meas.algorithms.detection.SourceDetectionTask.updatePeaks	(	self,
		fpSet,
		image,
		threshold
	)

Update the Peaks in a FootprintSet by detecting new Footprints and
Peaks in an image and using the new Peaks instead of the old ones.

Parameters
----------
fpSet : `afw.detection.FootprintSet`
    Set of Footprints whose Peaks should be updated.
image : `afw.image.MaskedImage`
    Image to detect new Footprints and Peak in.
threshold : `afw.detection.Threshold`
    Threshold object for detection.

Input Footprints with fewer Peaks than self.config.nPeaksMaxSimple
are not modified, and if no new Peaks are detected in an input
Footprint, the brightest original Peak in that Footprint is kept.

Definition at line 832 of file detection.py.

     def updatePeaks(self, fpSet, image, threshold):
         """Update the Peaks in a FootprintSet by detecting new Footprints and
         Peaks in an image and using the new Peaks instead of the old ones.
 
         Parameters
         ----------
         fpSet : `afw.detection.FootprintSet`
             Set of Footprints whose Peaks should be updated.
         image : `afw.image.MaskedImage`
             Image to detect new Footprints and Peak in.
         threshold : `afw.detection.Threshold`
             Threshold object for detection.
 
         Input Footprints with fewer Peaks than self.config.nPeaksMaxSimple
         are not modified, and if no new Peaks are detected in an input
         Footprint, the brightest original Peak in that Footprint is kept.
         """
         for footprint in fpSet.getFootprints():
             oldPeaks = footprint.getPeaks()
             if len(oldPeaks) <= self.config.nPeaksMaxSimple:
                 continue
             # We detect a new FootprintSet within each non-simple Footprint's
             # bbox to avoid a big O(N^2) comparison between the two sets of
             # Footprints.
             sub = image.Factory(image, footprint.getBBox())
             fpSetForPeaks = afwDet.FootprintSet(
                 sub,
                 threshold,
                 "",  # don't set a mask plane
                 self.config.minPixels
             )
             newPeaks = afwDet.PeakCatalog(oldPeaks.getTable())
             for fpForPeaks in fpSetForPeaks.getFootprints():
                 for peak in fpForPeaks.getPeaks():
                     if footprint.contains(peak.getI()):
                         newPeaks.append(peak)
             if len(newPeaks) > 0:
                 del oldPeaks[:]
                 oldPeaks.extend(newPeaks)
             else:
                 del oldPeaks[1:]
 

Member Data Documentation

◆ ConfigClass

lsst.meas.algorithms.detection.SourceDetectionTask.ConfigClass = SourceDetectionConfig

static

Definition at line 262 of file detection.py.

◆ makeSourceCatalog

def lsst.meas.algorithms.detection.SourceDetectionTask.makeSourceCatalog = run

static

An alias for run.

Deprecated:: Remove this alias after checking for where it's used

Definition at line 358 of file detection.py.

◆ negativeFlagKey

lsst.meas.algorithms.detection.SourceDetectionTask.negativeFlagKey

Definition at line 280 of file detection.py.

The documentation for this class was generated from the following file:

/j/snowflake/release/lsstsw/stack/Linux64/meas_algorithms/19.0.0-2-g3d9e4fb2+11/python/lsst/meas/algorithms/detection.py

Public Member Functions

Static Public Member Functions

Public Attributes

Static Public Attributes

Detailed Description

Contents

Description

Task initialisation

Invoking the Task

Configuration parameters

Debug variables

A complete example of using SourceDetectionTask

Constructor & Destructor Documentation

◆ __init__()

Member Function Documentation

◆ applyTempLocalBackground()

◆ applyThreshold()

◆ calculateKernelSize()

◆ clearMask()

◆ clearUnwantedResults()

◆ convolveImage()

◆ detectFootprints()

◆ display()

◆ finalizeFootprints()

◆ getPsf()

◆ makeThreshold()

◆ reEstimateBackground()

◆ run()

◆ setEdgeBits()

◆ tempWideBackgroundContext()

◆ updatePeaks()

Member Data Documentation

◆ ConfigClass

◆ makeSourceCatalog

◆ negativeFlagKey

◆ init()